Context
A multi-year ERP transformation creates an extended period in which legacy and target systems both matter. Downstream teams still need reliable data while sources, contracts, and operating models change.
Delivery
The work uses Databricks, Delta Live Tables, Delta Lake, Kafka, and Terraform to support a governed coexistence layer between Atlas and SAP. Shared Commerce models, quality controls, and secure delivery provide a stable boundary for downstream consumers.
Engineering decisions
- Model coexistence explicitly instead of hiding it behind ad hoc transformations.
- Protect shared Commerce contracts while source systems evolve.
- Place quality controls at the boundaries where legacy and target data meet.
- Keep infrastructure and pipeline delivery reproducible through Terraform.
Outcome
The coexistence layer supports a controlled legacy-to-SAP transition while maintaining shared models, quality expectations, and secure data delivery.